Bearing Fault Diagnosis with Impulsive Noise Based on EMD and Cyclic Correntropy

Abstract

Periodic pulses are an important fault feature of rolling bearings, so the ability to accurately and efficiently identify pulse components is important for bearing fault diagnosis. Due to the complicated wheel-rail contact relationship in actual train operation, it often generates many impulse noises which similar to the fault signal structure. Unfortunately, spectral kurtosis (SK) methods often fail to effectively diagnose under impulse noise. In order to solve this problem, this paper proposes a bearing fault diagnosis method based on Empirical Mode Decomposition (EMD) and cyclic correntropy (CCE) function. Compared with the SK method, the method proposed in this paper can effectively suppress the influence of impulse noise. Moreover, this paper also proposes a fault diagnosis evaluation index \( KR_{s} \) to quantitatively compare the diagnostic effects of different methods. Simulations and real data of the train axle are utilized to demonstrate the feasibility and effectiveness of the proposed method and index.